###
DOI:
计算机系统应用英文版:2016,25(2):162-166
本文二维码信息
码上扫一扫!
新的轨道特征提取算法
(上海工程技术大学 城市轨道交通学院, 上海 201620)
Extracting Algorithmic of Train Geometry Characteristic
(College of Urban Railway Transportation, Shanghai University of Engineering Science, Shanghai 201620, China)
摘要
图/表
参考文献
相似文献
本文已被:浏览 1164次   下载 2162
Received:May 26, 2015    Revised:July 14, 2015
中文摘要: 轨道沉降是影响列车快捷、安全运行的重要原因.提取出轨道部件稳定的目标特征,为后续SLAM算法提供轨道区域匹配点.通过研究轨道组成及部件,分析出能提取出稳定目标特征的轨道部件.在CCD获取图像后,对图像均衡化、增强与洞形填充以完成钢轨和扣件轮廓的初识别.二值膨胀生成线性0、90度结构,水平、垂直闭关运算,生成成对元素,完成钢轨和扣件初步轮廓提取.对处理后的图像采用改进的Canny算子进行边缘检测,最后根据实验进一步证明了改进后的Canny在直线特征检测上面具有速度快,降低内存等优点,从而会后续DSP硬件开发上节约Flash空间,对整个课题具有重大的理论支撑意义.
Abstract:Train settlement is an important reason for affecting the speed and safe operation of train, meanwhile the changing of train axle box acceleration caused by track irregularities can reflect the status information of track irregularity. Extracting the steady target characteristics of pathway can provide the matching point of orbit for subsequent SLAM algorithm. Through the analysis of components of the pathway, the steady target characteristic of pathway parts could be extracted. Image equalization, enhancing and filling in with hole is to be used to complete the rail and fasteners outline at the beginning of the recognition. Binary expansion and generating linear structure of 0 or 90 degree, horizontal and vertical closing-opening operation, generating elements in pairs is to be used to complete the rail and fasteners initial contour extraction. According to the experiments it will prove that the improved RHT on linear feature extraction can increase the speed of PC and reduce the memory space, which can reduce the Flash on the subsequent DSP hardware development. It has great theoretical significance in supporting for the whole project.
文章编号:     中图分类号:    文献标志码:
基金项目:国家自然科学基金(51405287);上海市科委重点支撑项目(13510501300);上海市研究生教育创新计划学位点引导布局与建设培育项目(13sc002);上海工程技术大学创新项目(14KY1008)
引用文本:
王昆,郑树彬.新的轨道特征提取算法.计算机系统应用,2016,25(2):162-166
WANG Kun,ZHENG Shu-Bin.Extracting Algorithmic of Train Geometry Characteristic.COMPUTER SYSTEMS APPLICATIONS,2016,25(2):162-166